| Force feedback is widely assumed to enhance performance in robotic surgery, but its benefits have not yet been systematically assessed. Further, the cost of force sensors that allow force feedback is prohibitive, due to the stringent specifications imposed by the surgical environment. We address both sides of this cost benefit analysis of force feedback by investigating the specific benefits allowed by force feedback and by proposing a novel force sensor design.;We demonstrate the benefit of force feedback in surgery through a series of psychophysical experiments. By investigating performance on tasks with and without force feedback, we find that the primary benefit of force feedback is that interaction forces are reduced. This results in an increase in patient safety, because high forces correlate directly with tissue trauma. Two mechanisms enable force reduction and other benefits: (1) force feedback transforms environmental interaction forces into mechanical constraints and (2) forces act as a source of information. Mechanical constraints passively reduce intrusions into environmental structures (and, thereby, forces) due to the interactions between hand compliance and environment stiffness. Because this benefit is passive, it happens without a cognitive response. Accordingly, these benefits occur instantaneously, on the time scale of mechanical interactions. Force feedback also allows additional manipulation strategies that take advantage of these physical interactions, potentially reducing mental workload of the surgeon. We also find that force feedback provides information to the surgeon. While the number of ways that forces can potentially inform surgeons is large, the interaction between training and other sources of information is complex. We find that training is necessary, in some contexts, to take full advantage of the informational benefits of force feedback in surgery.;We propose a three axis force sensor design using strain gages and Shape Deposition Manufacturing (SDM), where components are embedded inside a pourable epoxy. The resulting sensor (+/-2 N range, 0.15 RMS calibration error, 0.15 N drift over five minutes) is small enough for minimally invasive surgery, waterproof, and temperature insensitive. Adapting the SDM design for mass production is straightforward (sensors are cast inreusable molds), resulting in high performance, low cost, disposable sensors. |
